High Frequency Trading (HFT)

High frequency trading, usually shortened to HFT, is a style of automated trading in which computer programs buy and sell securities at extremely high speed and volume, typically holding positions for fractions of a second to a few minutes. Instead of a person clicking buy or sell, the decisions and order placement are done entirely by software following pre-set rules.

The core idea is to exploit very small, very short-lived edges many thousands of times a day. Common approaches include market making, where the firm continuously posts both buy and sell orders and profits from the tiny gap between them (the bid-ask spread), and statistical arbitrage, where algorithms spot small pricing differences between related instruments — say, the same stock trading on two exchanges — and trade to capture them before they disappear. Because each individual trade earns a sliver of profit, HFT firms depend on doing an enormous number of trades and on being faster than competing algorithms.

Speed is the whole game, which is why HFT firms invest heavily in infrastructure: locating their servers physically inside or near an exchange's data center (co-location) to shave microseconds off order transmission, using specialized fiber or microwave data links, and writing highly optimized code. A delay of even a few milliseconds can mean the difference between capturing a price edge and missing it entirely.

The nuance that trips people up is that HFT is not a single strategy, it is a category of execution style defined by speed and automation. Two HFT firms can be doing completely different things — one providing liquidity as a market maker, another trying to detect and trade ahead of large orders. This is also why HFT is controversial: supporters say it narrows spreads and adds liquidity, while critics argue certain tactics can disadvantage slower traders or add fragile, fast-moving volatility to markets.